A Case Study of the Effect of the Observation Time in 3D-Var on Near-Surface Wind Forecasts in a Limited Region

Abstract

The relationships between the prediction of near-surface winds and the corresponding time of observations in eastern China were explored using the Advanced Weather Research and Forecasting (WRF) model and the three-dimensional variational (3D-Var) scheme in the gridpoint statistical interpolation (GSI) system. A series of one-month experiments was conducted in January 2018 with different time window configurations from 0.01 to 3.0 h. The relationship between the wind observation time and the model forecast was non-linear. An observational time closer to the initial time in the model usually have greater impact on the prediction of near-surface wind speeds. Observations in the 0.4−0.8 h time window associated with abnormally high with large near-surface wind speeds provide a negative impact. The predictions improved at a much smaller rate when the time window was increased from 0.8 to 3.0 h. No significant difference was seen as the time window increased in wind direction predictions, even with large wind increments. The optimum configuration of the time window in the GSI 3D-Var system for predicting near-surface winds should therefore be 0.2 or 0.4 h. A better understanding of the relationships between the observations and the predictions will help select more effective observations when using the 3D-Var scheme. (Citation: Yang, J., M. Shao, Q. Wang, and X. Yang, 2019: A case study of the effect of the observation time in 3D-Var on near-surface wind forecasts in a limited region. SOLA, 15, 172− 177, doi:10.2151/sola.2019-031.)